Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A processing method for image tensor data

A processing method and data technology, applied in the field of data processing, can solve problems such as high computational complexity, unsuitable RBM algorithm, and multiple storage spaces

Active Publication Date: 2021-04-20
BEIJING UNIV OF TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, more storage space and higher computational complexity are required, which makes the RBM algorithm unsuitable for applications on ordinary devices or high-dimensional data.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A processing method for image tensor data
  • A processing method for image tensor data
  • A processing method for image tensor data

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0010]This method of image tension data is introduced into a restricted Bolzman TtrBM model with a Tensor Train structure. The input and output data are tested, and the weight of the intermediate layer is also used. Indicates that the restriction weight has the structure of Tensor Train; the number of free parameters in the intermediate layer is adjusted by adjusting the tensor TRAIN decomposition; the rank of the TT decomposition is indicated by different characteristics.

[0011]The model input and output data of the present invention are represented by tensile, and the weight of the intermediate layer is also indicated by tensive, in order to reduce the amount of weight of the intermediate layer, the restriction weight has Tensor Train, by adjusting the tensile TRAIN The number of free parameters in the intermediate layer is decomposed, and the number of free parameters of the weight layer increases linearly with the dimension of the sample data, which greatly reduces the number of ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a processing method for image tensor data, which can greatly reduce the number of free parameters in a model, has flexible restrictions on a weight layer, and is applicable to image tensor data of any order. This method of image tensor data processing introduces a restricted Boltzmann machine TTRBM model with a tensor train structure. The input and output data of this method are represented by tensors, and the weights of the middle layer are also represented by tensors. Indicates that the restricted weight has the structure of Tensor Train; the number of free parameters in the intermediate layer is controlled by adjusting the rank of tensor Train decomposition; the rank of TT decomposition is adjusted, and different features of the same size are represented.

Description

Technical field[0001]The present invention relates to the technical field of data processing, and more particularly to a processing method of image tensile data, which can be directly applied to an image tensus data of any of the order.Background technique[0002]Restricted Boltzmman Machine (RBM) is a two-layer neural network consisting of visible layers and hidden layers due to its strong features representing the ability, and is widely used in pattern identification and machine learning. The visible layers and hidden layer data in the traditional RBM are represented in vector form.[0003]However, data from the actual life today often has high dimensional characteristics. In order to apply RBM on these high dimensional data, the common method is to quantify data, and the quantization process often destroy the internal structure in high dimensional data, resulting in a loss of important associated information, or generating a dimension disaster. In addition, RBM is a fully connected n...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08G06V10/40G06V10/50
CPCG06N3/08G06N3/045G06F18/2133G06V40/169G06V10/40G06V10/50G06V10/82G06N3/047G06F18/29G06N20/10G06F17/18G06T3/4046G06T3/4053
Inventor 孙艳丰句福娇
Owner BEIJING UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com